The Effects of Land-Atmosphere Interactions on Convection Initiation and Quantitative Precipitation Forecasts during the International H2O Project (IHOP_2002)

Abstract

The International H₂O Project (IHOP_2002) was a field study conducted over the Southern Great Plains (SGP) during the summer of 2002. The main goal of IHOP_2002 was to study water vapor distribution in the atmosphere in order to improve predictions of warm season rainfall over the SGP. This research is aimed toward studying the effects of interactions between the land and atmosphere on mesoscale model simulations of convection and precipitation. The convection case examined during this study occurred from 17-20 June 2002 during IHOP_2002. A combination of synoptic forcing and surface features contributed to the development and continuance of convection during this time period. An observational analysis of this case showed evidence of surface dryline boundaries, accompanied by temperature and moisture gradients. Gradients of surface sensible and latent heat fluxes across the boundaries were also seen throughout the time period. Convection formed along the boundaries, suggesting that the land-surface features are important in the development of deep moist convection, even for cases with strong synoptic forcing.

Two sets of sensitivity studies were conducted for this case using the MM5 model. The first involved testing the sensitivity of different surface data assimilation schemes on simulations of convection. A Control Simulation was performed using standard MM5 data assimilation, followed by an Experimental Simulation using a Flux-Adjusting Data Assimilation System (FASDAS). Significant differences were seen between these two simulations, especially in regards to accumulated precipitation. The Control Simulation generated very little precipitation, while the Experimental Simulation produced significantly more precipitation, and overall appeared to match more closely with observations than the Control. The Experimental Simulation also produced more varied heat flux fields, setting up strong gradients which may have led to the enhanced precipitation. The Experimental Simulation also overall matched observed heat fluxes than the Control Simulation. This study suggests that improvements in surface data assimilation may drastically improve predictions of convection and precipitation.

The second study involved testing the effects of using two different coupled land-surface models (LSM) and planetary boundary layer (PBL) schemes on simulations of convection. The same convection case (17-20 June 2002) was used for this study. The first simulation used the Noah LSM coupled to the MRF PBL, and the second used the Pleim-Xiu (PX) LSM coupled to the ACM PBL. The precipitation fields generated by both simulations were very similar to each other; however they differed from the observations in terms of accumulations and timing. Significant differences were seen in the evolution of soil moisture and temperature between the two simulations, which was to be expected due to the different soil layers in each LSM used. There was a more immediate soil-precipitation feedback in the PX run, where the top soil layer was closer to the surface. A similar feedback could be seen in the heat flux fields. Results of this study provide evidence of the importance of the land-surface in mesoscale modeling, while also suggesting that improvements are needed in modeling land-surface processes.